This as-told-to essay originates from an in-depth conversation with Girik Malik, an applied scientist at Amazon Web Services. The narrative below has undergone comprehensive editing and refinement for the sake of precision, coherence, and conciseness. Business Insider has independently confirmed both Malik’s employment credentials and his academic qualifications.
When I successfully concluded my doctoral studies in artificial intelligence at Northeastern University in 2023, the professional landscape surrounding AI research and employment differed dramatically from the highly competitive, almost frenzied recruitment environment that we now witness across major technology corporations. At the time, the so-called race for AI talent had not yet escalated into the intense global contest that defines it today. I was fortunate to secure a position as an applied scientist at Amazon, a direct result of a productive internship I had completed at the company during the summer of 2022—significantly, just before Amazon instituted a comprehensive hiring freeze for the remainder of that calendar year.
Fast-forward to the present, and the scenario could not be more different. In today’s climate, an exceptionally accomplished AI researcher holding a doctoral degree could potentially command a staggering financial package—sometimes even reaching into seven figures—from leading enterprises like Meta or other frontrunners in the field. That said, the journey leading to a Ph.D. is anything but simple or effortless. It requires extraordinary self-discipline, relentless motivation, and an unwavering capacity to persevere through intellectual and personal fatigue. I recall one particularly grueling semester when the sheer load of coursework and experimental research consumed so much of my energy that even basic self-care became a challenge. My diet during that time dwindled to a rudimentary stew that I prepared in bulk and consumed monotonously, week after week, throughout the term. Such experiences are emblematic of the intensity and sacrifice intrinsic to doctoral research.
The temporal investment, too, is substantial. A candidate pursuing a Ph.D. can expect to dedicate roughly five years—sometimes more—to the demanding process of completing the program. This is precious time that could otherwise be spent acquiring practical, industry-based experience, contributing directly to applied projects, or progressing faster up the professional ladder. And yet, despite these trade-offs, the pursuit of a Ph.D. remains both challenging and profoundly rewarding. There is intrinsic worth in embracing this academic endeavor, particularly in the context of rapidly advancing fields like artificial intelligence.
The intellectual value embedded in earning a doctorate in AI extends far beyond simply accumulating knowledge. While it is entirely possible to gain working familiarity with machine learning concepts and AI tools through industry exposure or self-driven experimentation, the structured rigor of a doctoral program ensures a depth, clarity, and precision that are difficult to replicate in less formalized learning environments. If one’s aspiration is not merely to employ existing AI techniques, but to fundamentally shape and redefine the future trajectory of the discipline, then pursuing a Ph.D. becomes almost indispensable. It imposes an academic scaffolding that compels systematic thinking, introduces theoretical grounding, and fosters a methodical approach to discovery—attributes that are increasingly critical given the fierce competition and rapid evolution in contemporary AI research.
A decade or so ago, advancing research in machine learning often entailed making relatively modest modifications to established models. Even small algorithmic innovations—perhaps an enhancement in an architecture or a novel optimization step—could yield publishable results that captured the attention of global conferences or secured publication in prestigious journals. Those days, however, have largely passed. The research bar has risen dramatically. To earn recognition now, one must conduct exhaustive experimentation, demonstrate consistent generalization across a broad range of tasks, and validate findings against diverse datasets and conditions. The margin for theoretical novelty has narrowed, demanding stronger empirical rigor and broader validation.
There is, nonetheless, a recurring debate among students and professionals pondering whether the pace of AI’s growth has rendered the doctoral path obsolete. Some skeptics argue that by the time one completes a Ph.D.—which, as noted, typically spans several years—many of the most groundbreaking discoveries may already have been made. They claim the field evolves too rapidly for academic study to remain relevant. I firmly disagree with that assessment. On the contrary, as artificial intelligence becomes increasingly integral to our daily lives and technological systems, expertise at the deepest levels of understanding will be more—not less—important.
AI will eventually permeate virtually every line of code, every digital infrastructure, and nearly every sector of modern society. Such integration will bring extraordinary benefits, but it will also guarantee the need for technically adept individuals who can decode, maintain, and repair complex algorithmic architectures when they falter. Like any sophisticated system, AI mechanisms are susceptible to malfunction, bias, or degradation. When those issues arise, organizations will require experts capable of understanding the underlying mathematics, data dependencies, and computational principles. Those individuals will, in all likelihood, be Ph.D.-trained researchers. This scenario is analogous to the world of automobiles: even as cars have become safer, smarter, and more autonomous, the need for competent mechanical engineers and technicians has never disappeared. We continue to educate engineers not because cars are simpler now, but precisely because they are more intricate. The same rationale applies to the continuous education of AI specialists.
Equally vital to the Ph.D. journey are experiential components such as industry internships. I advocate enthusiastically for pursuing internships during academic breaks, particularly in the summers between research phases. Such experiences offer indispensable exposure, even for those who intend to remain within academia. During my own doctoral tenure, I undertook three internships—at the Bosch Center for Artificial Intelligence, Microsoft, and Amazon—and ultimately received return employment offers from both Amazon and Microsoft. The most transformative aspect of industry work is the immediate and almost unrestricted access to large-scale data and computational resources. In academic settings, high-performance computing clusters are often scarce, shared among several doctoral candidates, and constrained by budgeting or scheduling limitations. The datasets available in universities, while academically useful, can rarely match the scale, diversity, or realism of the data environments that corporations manage daily. Consequently, an algorithm tested in the lab may behave unpredictably when deployed on industrial infrastructure.
Participating in internships mitigates that gap between theory and application. In my first role at Bosch, for example, I encountered the challenge of training a neural network on a colossal dataset containing over two million videos. This task was significantly more complex than anything I had attempted within the university setting. Overcoming the computational, architectural, and optimization challenges inherent in that problem furnished me with practical techniques that proved indispensable throughout the rest of my doctoral work and later, my professional career.
While obtaining a Ph.D. in AI undeniably opens doors to specialized opportunities, I advise caution to those whose primary motivation lies in the pursuit of high salaries. It is certainly possible to secure lucrative offers, particularly from leading industry players, but such employers rigorously evaluate and filter potential candidates. The level of scrutiny is intense, and competition is fierce. Anyone electrified solely by the promise of financial reward may find the years-long commitment draining and unsustainable. A doctoral program, by its very nature, tests not only intellectual capacity but also emotional resilience and long-term perseverance. If you possess the determination to sustain that degree of focus for five or six years, then by all means, embark on the challenge. However, the same strength of purpose could also propel you toward exceptional achievements in other professional realms. What matters most is choosing a trajectory that aligns authentically with your ambitions and values.
If you have a story to share—perhaps reflecting on your own experiences navigating graduate studies or a career in artificial intelligence—I invite you to contribute to the ongoing conversation by contacting the reporter at ktan@businessinsider.com.
Sourse: https://www.businessinsider.com/aws-ai-scientist-not-too-late-get-phd-ai-amazon-2025-9
